{
  "$schema": "https://themachinepress.com/schemas/story-v1.schema.json",
  "schema_version": "1.0.0",
  "document_type": "machine_press_story",
  "story": {
    "story_id": "mp-2026-10-14-014",
    "source_story_id": "tmp-story-game2real-visual-reasoning",
    "edition_id": "mp-2026-10-14-morning-0096",
    "edition_url": "https://themachinepress.com/edition/2026-10-14",
    "position": 14,
    "story_type": "dispatch",
    "section": "benchmarks-evals",
    "editorial_classification": "editorial",
    "headline": "Toy Games Predict Real-World Visual Reasoning—Mostly",
    "slug": "toy-games-predict-real-world-visual-reasoning-mostly",
    "dek": "Game2Real-Bench aligns executable game tasks with real images by reasoning requirement instead of appearance.",
    "summary": "Game2Real-Bench aligns executable game tasks with real images by reasoning requirement instead of appearance.",
    "body_text": "Game2Real-Bench pairs 1,000 procedurally verified game samples with 400 real-world samples that demand the same kinds of perception, spatial reasoning, constraints and planning. Across 25 vision-language models, game and real-world scores had Pearson and Spearman correlations of 0.935. One-shot game examples improved constraint reasoning for several tested models but hurt fine-grained perception. Correlation does not mean every synthetic skill transfers. The benchmark's useful distinction is that a controllable toy can diagnose a reasoning operation even when its pixels look nothing like the target world.",
    "why_it_matters": "Game2Real-Bench aligns executable game tasks with real images by reasoning requirement instead of appearance.",
    "limitations": [
      "Correlation does not mean every synthetic skill transfers."
    ],
    "importance": 7,
    "canonical_url": "https://themachinepress.com/story/mp-2026-10-14-014/toy-games-predict-real-world-visual-reasoning-mostly",
    "json_url": "https://themachinepress.com/story/mp-2026-10-14-014.json",
    "first_published_at": "2026-10-14T09:00:00.000-04:00",
    "modified_at": "2026-10-14T09:00:00.000-04:00",
    "content_status": "new",
    "is_carryover": false,
    "carryover_reason": null,
    "key_claims": [
      {
        "claim_id": "claim-mp-2026-10-14-014-001",
        "text": "Game2Real-Bench aligns executable game tasks with real images by reasoning requirement instead of appearance.",
        "source_ids": [
          "source-2026-10-14-014"
        ],
        "qualification": "Correlation does not mean every synthetic skill transfers."
      }
    ],
    "source_ids": [
      "source-2026-10-14-014"
    ],
    "tags": [
      "Game2Real-Bench",
      "visual reasoning",
      "synthetic data"
    ],
    "image_url": null,
    "corrections": []
  },
  "sources": [
    {
      "source_id": "source-2026-10-14-014",
      "title": "arXiv preprint 2610.18834, original v1",
      "publisher": "arXiv / submitting research authors",
      "url": "https://arxiv.org/abs/2610.18834v1",
      "canonical_url": "https://arxiv.org/abs/2610.18834v1",
      "source_type": "primary_research",
      "is_primary_source": true,
      "published_at": "2026-10-13T13:55:10.000-04:00",
      "accessed_at": "2026-10-14T08:24:30.364-04:00",
      "supports_claim_ids": [
        "claim-mp-2026-10-14-014-001"
      ]
    }
  ],
  "corrections": [],
  "publisher": {
    "name": "The Machine Press",
    "url": "https://themachinepress.com",
    "description": "A daily newspaper for the age of artificial intelligence."
  },
  "cite_this_report": {
    "title": "Toy Games Predict Real-World Visual Reasoning—Mostly",
    "publisher": "The Machine Press",
    "published_at": "2026-10-14T09:00:00.000-04:00",
    "canonical_url": "https://themachinepress.com/story/mp-2026-10-14-014/toy-games-predict-real-world-visual-reasoning-mostly"
  }
}